FEATURE SELECTION AND CLASSIFICATION OF METABOLOMIC DATA USING SUPPORT VECTOR MACHINES

摘要

Over the past few years there has been an explosion of biological data available for exploratory analysis. The main task of data analysis is to extract meaningful information in a way that facilitates the understanding of the complex biological processes. In order to do this, algorithms and techniques have to be developed that can be trained to learn rules and form patterns from the available data sets and then apply these rules to analyse new data. In computing science terminology this is known as machine learning. In this paper, the applicability of one such machine learning technique, namely 'support vector machines' to analyze and classify metabolomic data is explored. The paper also explores some of the feature selection algorithms which help determine important biomarkers or metabolites in data sets.